November 6, 2020, ainerd
AI Is Making Magic With Music
How Does Ai Help Musicians Write Music?
Artificial intelligence is changing the way artists think about music production, and predicts that creative musicians will have to work creatively with machines. AI – generated songs are simple and sometimes not useful at all, musicians should not worry about losing their job. While the idea that machines will produce music may seem daunting to musicians, the truth is that artificial intelligence can provide them with new creative ideas. When machines compose songs, implications can run the gamut from what makes a person unique to the possibility of human-machine interaction.
This can be a sore point, but AI can be a powerful tool that can stimulate creativity and allow composers and producers to remain in a state of creative flow. While some areas of music production, such as music, may be under threat, creative AI is likely to be seen by songwriters and publishers as a tool to support the otherwise difficult creative process.
Even if AGI is put aside and we focus on the current AI songwriting tools aimed at the music industry, these tools could open up new opportunities for musicians without replacing them.
Popgun hopes that their instruments will complement human musicians and spark new creativity and collaboration, rather than stealing opportunities from musicians. Although audio mastering still has a creative component, AI could make it easier for artists not to master their songs, even if they prefer to rely on people to do the work. This is how we see the role of AI music in the creative ecosystem: rather than stealing the jobs of real artists, it could offer musicians enormous benefits: musicians can now spend enormous amounts of time and money investing their abundant talent for expression in songwriting, symphony, and storymaking. AI in music takes jobs away, but it doesn’t steal real musicians’ jobs at all.
The bottom line is that most AI music systems are really good at composing and producing instruments, but they don’t understand the song structure yet. It’s not about making great music. This is not easy – but to familiarize oneself with the AI models that generate the melodies.
Musicians who turn to AI can use it as an inspiring tool because it has the ability to crank up an entire song. Musicians who have difficulty remembering their own ideas will love Humtap because they can simply hum a melody and the app automatically generates the entire song with different instruments. You can record a guitar track from a musician you really admire (say, he mixes a song with a bunch of different guitar tracks from different genres) and ask the Flow Machines to map the track. He can do it for you in a few seconds or even less than a minute, and he can do it in less than 30 seconds.
Google introduced it this year and gets you to play piano tones, and then the AI responds with its own tune. Suppose you feed it the musically plausible next beat predicted by Google’s free app Continue of Watson’s Beat.
The team fed its neural network with huge chunks of music data to analyze, and the AI learned by writing original music, for example. Crescendo uses the musical expertise that enables it to give feedback and point out mistakes. The musician plays the piece, the AI learns from it and she also learns by writing her own melody.
This approach represents a significant step forward for composers who also consider the programming tools as songwriting credits and royalties. In his presentation at FutureFest, Paul Mason postulated that musical artificial intelligence will produce music that we cannot think, but could also produce music that we cannot understand. The future may require musicians to have more control over their own music than ever before and to use their knowledge of the world around them to determine what kind of music needs to be created.
As someone who is passionate about AI in music, Herndon is worried about harming the musicians who train the machines and make them effectively what they are. He fears that AI’s music could create an ad hoc effect, with large companies profiting from lax intellectual property laws while independent musicians remain unpaid and unrecognised. Amper Music CEO Drew Silverstein claims the results will benefit people and boldly predicts that AI Music will eventually become indistinguishable from human-made music.
That doesn’t sound like such a grim prospect, but what if AI tools can’t understand context or purpose? He talks about using AI to work on apps that can create music for composers plagued by writing blocks. It is like having a composer who has the ability to read 1000 pieces of music and then try to invent something similar.
It is therefore important to feed music knowledge, representation and music theory into the algorithm and combine it with machine learning.